Robotic radiotherapy using extensively non-coplanar beams has been shown effective to significantly improve radiation therapy dosimetry that leads to improved treatment outcome. However, current implementation of this technique by CyberKnife is inefficient and not optimal dosimetrically. This has severely limited both the number of patients eligible for robotic radiotherapy and the achievable clinical outcome for those who have been treated. In order to overcome these limitations, we propose to develop a novel robotic radiotherapy system that can efficiently utilize the full potential of the non-coplanar delivery space to treat the majority of radiotherapy patients. Innovation: The proposed system is highly innovative in the following aspect: 1) Integrated beam orientation and fluence optimization. 2) Significantly more compact linac to allow posterior beams. 3) Flexible field sizes and MLC resolution to efficiently treat most target sizes. 4) Integrated volumetric imaging system. This project is proposed to design a hardware platform materializing such robotic radiotherapy system. In order to reduce the gantry size, both the linac length and the distance between the source and the MLC need to be significantly reduced. We propose to design a new 2 MV source to reduce linac length and provide the required dose rate for treatment. The physical MLC leaf thickness cannot be substantially thinner than 1 mm. To achieve a high MLC resolution at the treatment distance, a spacer is used in CyberKnife between the primary collimator and the MLC, increasing the gantry dimension. We propose to eliminate the spacer but vary the focus-to-tumor distances (FTD) to achieve desired field size and MLC resolution. This requires optimization in an enormous solution space, a capacity uniquely demonstrated by the 4p algorithm. Volumetric imaging has been an indispensable component of modern radiotherapy but unfortunately missing from existing robotic systems. The proposed new linac will be able to deliver kV imaging beams from the same 2 MV linac, which in combination with gantry or couch mounted imagers will allow volumetric imaging for more precise tumor targeting. Aims: 1: Prototypical design of the accelerator to produce 2 MV X-rays 2: Design incorporated imaging system 3: Develop a conceptual design for the entire clinical system Impact: Success of the Phase I project would lead to the design of the first 2 MV linear accelerator capable of producing a competitively high dose rate of >800 cGy/min at 100 cm and kV imaging beams for image guided radiotherapy. This paves the technical path to a new robotic radiotherapy system delivering radiation plans with dose conformality surpassing existing X-ray platforms. More importantly, the significantly increased field size, throughput and the volumetric imaging capacity would allow the new robotic system to compete for a much larger market, including that for conventional linacs, than the niche market CyberKnife currently commands.